Formation of aroma compounds from ribose and cysteine during the Maillard reaction

Article abstract of DOI:10.1021/jf026123f

The headspace volatiles produced from a phosphate-buffered solution (pH 5) of cysteine and a 1 + 1 mixture of ribose and [¹³C₅]ribose, heated at 95 °C for 4 h, were examined by headspace SPME in combination with GC-MS. MS data indicated that fragmentation of ribose did not play a significant role in the formation of the sulfur aroma compounds 2-methyl-3-furanthiol, 2-furfurylthiol, and 3-mercapto-2-pentanone in which the carbon skeleton of ribose remained intact. The methylfuran moiety of 2-methyl-3-(methylthio)furan originated from ribose, whereas the methylthio carbon atoms came partly from ribose and partly from cysteine. In 3-mercapto-2-butanone one carbon unit was split from the ribose chain. On the other hand, all carbon atoms in 3-thiophenethiol stemmed from cysteine. In another trial cysteine, 4-hydroxy-5-methyl-3(2H)-furanone and [¹³C₅]ribose were reacted under the same conditions. The resulting 2-methyl-3-furanthiol was mainly ¹³C5-labeled, suggesting that it stems from ribose and that 4-hydroxy-5-methyl-3(2H)-furanone is unimportant as an intermediate. Whereas 2-mercapto-3-pentanone was found unlabeled and hence originated from 4-hydroxy-5-methyl-3(2H)-furanone, its isomer 3-mercapto-2-pentanone was formed from both 4-hydroxy-5-methyl-3(2H)-furanone and ribose. A new reaction pathway from ribose via its 1,4-dideoxyosone is proposed, which explains both the formation of 2-methyl-3-furanthiol without 4-hydroxy-5-methyl-3(2H)-furanone as an intermediate and a new way to form 3-mercapto-2-pentanone.

Full text of DOI:10.1021/jf026123f

2714 J. Agric. Food Chem. 2003, 51, 2714−2721
Formation of Aroma Compounds from Ribose and Cysteine
during the Maillard Reaction
§
CHRISTOPH CERNY*,† AND TOMAS DAVIDEK
Nestle´ Research Center, Vers-chez-les-Blanc, P.O. Box 44, CH-1000 Lausanne 26, Switzerland, and
Centres R&D Nestle´ S.A.S., Amiens, P.O. Box 47, F-80800 Aubigny, France
The headspace volatiles produced from a phosphate-buffered solution (pH 5) of cysteine and a 1 +
1 mixture of ribose and [13C5]ribose, heated at 95 °C for 4 h, were examined by headspace SPME
in combination with GC-MS. MS data indicated that fragmentation of ribose did not play a significant
role in the formation of the sulfur aroma compounds 2-methyl-3-furanthiol, 2-furfurylthiol, and
3-mercapto-2-pentanone in which the carbon skeleton of ribose remained intact. The methylfuran
moiety of 2-methyl-3-(methylthio)furan originated from ribose, whereas the methylthio carbon atoms
came partly from ribose and partly from cysteine. In 3-mercapto-2-butanone one carbon unit was
split from the ribose chain. On the other hand, all carbon atoms in 3-thiophenethiol stemmed from
cysteine. In another trial cysteine, 4-hydroxy-5-methyl-3(2H)-furanone and [13C5]ribose were reacted
under the same conditions. The resulting 2-methyl-3-furanthiol was mainly 13C5-labeled, suggesting
that it stems from ribose and that 4-hydroxy-5-methyl-3(2H)-furanone is unimportant as an intermediate.
Whereas 2-mercapto-3-pentanone was found unlabeled and hence originated from 4-hydroxy-5-
methyl-3(2H)-furanone, its isomer 3-mercapto-2-pentanone was formed from both 4-hydroxy-5-methyl-
3(2H)-furanone and ribose. A new reaction pathway from ribose via its 1,4-dideoxyosone is proposed,
which explains both the formation of 2-methyl-3-furanthiol without 4-hydroxy-5-methyl-3(2H)-furanone
as an intermediate and a new way to form 3-mercapto-2-pentanone.
KEYWORDS: Maillard reaction; 2-methyl-3-furanthiol; 2-furfurylthiol; 2-mercapto-3-pentanone; 3-mercapto-
2-pentanone; ribose; [13C5]ribose; cysteine; 4-hydroxy-5-methyl-3(2H)-furanone; stable isotope labeling;
solid-phase microextraction; 1,4-dideoxyosone; meatlike flavor
INTRODUCTION
tanone, and 2-mercapto-3-pentanone (15, 16). Model experi-
ments with hydrogen sulfide and 2-furaldehyde, formed in the
Maillard reaction from pentoses via the 3-deoxyosone, showed
that this intermediate is important for the formation of 2-fur-
furylthiol (7, 17). Another Maillard reaction pathway from
pentoses via the 1-deoxyosone leads to 4-hydroxy-5-methyl-
3(2H)-furanone (18). 4-Hydroxy-5-methyl-3(2H)-furanone is
considered to be an important meat aroma precursor (19), and
model reactions of 4-hydroxy-5-methyl-3(2H)-furanone with
cysteine and hydrogen sulfide confirmed that 2-methyl-3-
furanthiol, 3-mercapto-2-pentanone, 2-mercapto-3-pentanone,
and 2-methyl-3-thiophenethiol are generated from it (15, 19-
21).
The Maillard reaction plays a substantial role in flavor
generation during the cooking of meat. Besides thiamine, the
precursors ribose and cysteine are important for meat aroma
and are consequently employed in the production of process
flavors (1). Ribose is frequently replaced by its less expensive
isomer xylose. Among other compounds, the thiols 2-methyl-
3-furanthiol, 2-furfurylthiol, and 3-mercapto-2-pentanone belong
to the most important aroma impact compounds formed during
the thermal reaction of ribose and cysteine. These molecules
are also found in cooked meat (2-4), as well as in commercial
meat flavorings (5), and contribute significantly to their aroma.
Numerous model reactions with ribose and cysteine have been
carried out to study meat flavor generation (6-14). Further
studies used hypothetical intermediates that could be formed
during these reactions. The diketones 2,3-butanedione and 2,3-
pentanedione seem to be intermediates in the formation of the
R-mercaptoketones 3-mercapto-2-butanone, 3-mercapto-2-pen-
Isotopically labeled compounds are very useful in aroma
research. Aroma compounds can be accurately quantified with
the help of their corresponding labeled compounds in isotope
dilution analysis (22). Labeled precursors in model reactions
are employed to elucidate formation pathways, for example, in
pyrazine formation (23). Recently the use of defined mixtures
of labeled and unlabeled precursors in model studies was
proposed to gain information on the fragmentation of this
precursor during aroma formation (24-26). This method allows,
for example, the determination of the extent to which a carbon
* Address correspondence to this author at Firmenich SA, P.O. Box 148,
CH-1217 Meyrin 2 Geneva, Switzerland (telephone +41-22-780 2216; fax
+41-22-780 2735; e-mail christoph.cerny@firmenich.com).
† Nestle´ Research Center, Vers-chez-les-Blanc.
§ Centres R&D Nestle´ S.A.S., Amiens.
10.1021/jf026123f CCC: $25.00 © 2003 American Chemical Society
Published on Web 03/29/2003

Aroma Compounds from Ribose and Cysteine
J. Agric. Food Chem., Vol. 51, No. 9, 2003 2715
Table 1. Model Reactionsa
amount (µmol)
A
B
C
D
E
cysteine
ribose
48.3 48.3 48.3 48.3 48.3
144.9
72.5
[13C ]ribose
72.5 72.5 72.5
5
4-hydroxy-5-methyl-3(2H)-furanone
2-furaldehyde
72.5
72.5
72.5
a Reaction in phosphate buffer (0.5 mol/L; pH 5.00) at 95 °C (4 h).
chain of a sugar precursor remains intact during the Maillard
reaction and hence gives insight into the significance of different
formation pathways. This approach was called “carbohydrate
module labeling” (25) or “carbon module labeling” (26),
respectively, and CAMOLA was used as acronym in both cases.
The objective of this study was to elucidate the formation
pathways for odorants formed from ribose and cysteine during
the Maillard reaction. The CAMOLA approach should explain
to what extent the ribose skeleton remains intact during the
formation of odorants and to what extent breakdown products
play a role. Additional experiments reacting a mixture of a
carbohydrate (ribose) and an intermediate [4-hydroxy-5-methyl-
3(2H)-furanone and 2-furaldehyde, respectively] with cysteine
should elucidate the role of the intermediate in the formation
of the aroma compounds. In this approach the intermediate and
the sugar contain different carbon isomer labels (13C and 12C),
which allows the actual contribution of the intermediate to the
formation of specific aroma compounds to be estimated.
Figure 1. Compounds formed from ribose and cysteine (numbering refers
to Table 2).
buffer at 95 °C for 4 h. These conditions are within the frame
of process flavor manufacturing conditions as defined by the
Council of Europe, whose guidelines limit the reaction tem-
perature to 180 °C, the reaction time to 24 h (for temperatures
e 110 °C), and the pH value to a maximum of 8.0 (27). A
similar reaction was carried out by replacing ribose with a
mixture (1:1) of unlabeled and 13C5-labeled ribose. Both reaction
mixtures were analyzed by SPME using an absorption time of
60 min and a PDMS-DVB fiber, followed by desorption and
GC-MS. 2-Methyl-3-furanthiol (7), 3-mercapto-2-pentanone (8),
2-furfurylthiol (9), and many other aroma compounds were
successfully detected with this method in both mixtures. The
structures of the identified volatiles are shown in Figure 1.
The mass spectra of the many volatiles isolated from the
ribose/[13C5]ribose/cysteine trials showed signals with mass
differences up to M+ + 10 as compared to the spectra of the
corresponding volatiles from the ribose/cysteine trial. These
signals correspond to compounds bearing 13C atoms originating
from [13C5]ribose. Table 2 lists the identified volatiles and the
proportion of their 13C-labeled isotope molecules. Values for
the M+ + 1 and M+ + 2 signals were corrected by subtracting
the naturally occurring percentages of 13C (1.10%), 33S (0.76%),
and 34S (4.20%). The loss of hydrogen frequently observed with
the molecular ion in EI-MS was also corrected in the labeled
molecular ions by the ratio (M+ - 1)/M+.
In the case of 2-methyl-3-furanthiol (7) an isotope peak m/z
119 appeared in the mass spectrum in addition to m/z 114 (M+).
The unlabeled and the 5-times-labeled 2-methyl-3-furanthiol
occurred at an approximate ratio of 1:1. Other labeling for
2-methyl-3-furanthiol (m/z 115-118) was practically not ob-
served, indicating that during 2-methyl-3-furanthiol formation
the ribose carbon skeleton remained intact. During studies on
several intermediates Hofmann and Schieberle found the highest
yields for 2-methyl-3-furanthiol in the reaction between hy-
droxyacetaldehyde and mercapto-2-propanone (28). In the
present study, however, pathways via ribose fragmentation were
not relevant.
MATERIALS AND METHODS
Chemicals. Ribose and cysteine were from Fluka (Buchs, Switzer-
land), 2-furaldehyde was from Aldrich (Buchs, Switzerland), dipotas-
sium hydrogenphosphate and potassium dihydrogenphosphate were
from Merck (Darmstadt, Germany), 4-hydroxy-5-methyl-3(2H)-fura-
none was from Advanced Biotech (Paterson, NJ), and [13C5]ribose (98%
enrichment) was from Cambridge Isotope Laboratories (Andover, MA).
All chemicals were of analytical grade.
Reactions. The reactants listed in Table 1 were dissolved in
potassium phosphate buffer (0.5 mol/L; pH 5.00) to a total amount of
500 mg. The solutions were filled into silanized 2 mL glass vials and
the septum-closed vials thermally treated for 4 h in a heated metal
block (Reacti-Therm, stirring/heating module, Pierce Chemical Co.,
Rockford, IL) at 95 °C. The molar ratio between cysteine and the other
precursors was 1:3, except for the reaction between cysteine and
4-hydroxy-5-methyl-3(2H)-furanone (2:3).
Analysis. All samples were analyzed in duplicate by headspace solid-
phase microextraction in combination with gas chromatography coupled
to mass spectrometry (HS-SPME-GC-MS). After at least 1 h of
equilibration at 20 °C, the fiber [polydimethylsiloxane-divinylbenzene
(PDMS-DVB), film thickness ) 65 µm, Supelco] was exposed for 60
min at 20 °C to the headspace above the samples in the glass vials
without agitation. After sampling, the SPME device was placed for 5
min in the GC injector, equipped with a 0.75 mm i.d. liner (Supelco),
and heated at 250 °C. GC-MS analyses were performed on a GC 6890A
coupled to an MSD 5973 (both Agilent, Palo Alto, CA) using an HP-
PONA column (50 m × 0.20 mm × 0.50 µm, Agilent). After insertion
of the SPME device into the injector, the oven temperature program
was started and the temperature raised at 6 °C/min from 35 to 240 °C
and held for 10 min isothermally. Mass spectra in the electron impact
mode (EI) were generated at 70 eV and at a scan range from m/z 28 to
350.
RESULTS AND DISCUSSION
2-Furfurylthiol (9) and 2-furaldehyde (6) showed similar
formation patterns with either only unlabeled or completely
labeled compound in a ratio of 1:1. These data support the
Reaction of Ribose, [13C5]Ribose, and Cysteine. Cysteine
and ribose (molar ratio 1:3) were reacted at pH 5 in phosphate

2716 J. Agric. Food Chem., Vol. 51, No. 9, 2003
Cerny and Davidek
Table 2. Proportion of Isotopomers from the Reaction between Ribose, [13C ]Ribose, and Cysteine
5
proportion of labeled carbon atoms in the moleculea (%)
+
no.
compoundb
m/z (M )
0c
1
2
3
4
5
6
10
1
2
3
4
5
6
7
8
furan
2-methylfuran
thiazole
2-methylthiophene
3-mercapto-2-butanone
2-furaldehyde
2-methyl-3-furanthiol
3-mercapto-2-pentanone
2-furfurylthiol
2-methyl-3-(methylthio)furan
3-thiophenethiol
2-methyl-3-thiophenethiol
bis(2-methyl-3-furyl) disulfide
(2-methyl-3-furyl) (2-oxo-3-pentyl) disulfided
bis(2-furfuryl) disulfide
68
82
85
98
104
96
114
118
114
128
116
130
226
230
226
49
51
56
45
49
47
49
49
48
36
95
44
26
27
28
0
0
0
0
0
0
0
0
0
1
1
2
6
0
1
0
1
1
0
0
1
2
1
1
1
0
1
14
0
0
50
9
0
44
7
0
0
1
1
0
11
1
39
44
3
50
5
3
2
2
2
1
10
3
45
46
47
47
36
9
10
11
12
13
14
15
13
2
0
0
0
24
48
49
54
0
0
0
22
20
18
3
0
0
+
+
a Values are corrected by subtracting the naturally occurring percentages of 13C (1.10%), 33S (0.76%), and 34S (4.20%) in M + 1 and M + 2. The loss of hydrogen
+
+
observed with the molecular ion in EI-MS was also corrected in the labeled molecular ions by the ratio (M − 1)/M . b Compounds were identified in the corresponding
reaction between ribose and cysteine by comparing the mass spectra and retention indices with those of authentic compounds analyzed in our laboratory or in literature
+
data. c Number of 13C atoms in the molecule. d Mass spectrum of the unlabeled compound m/z (%) 230 (M , 43), 187 (13), 145 (27), 114 (34), 113 (92), 85 (15), 81 (14),
43 (100).
formation of 2-furfurylthiol from ribose via 2-furaldehyde as
postulated by Silwar and Tressl (29).
The R-mercaptoketones 3-mercapto-2-butanone (5) and 3-mer-
capto-2-pentanone (8) were not generated from ribose fragments;
only completely unlabeled or completely 13C5-labeled com-
pounds in equal amounts were found. Obviously, 5 was formed
from ribose through the loss of one carbon atom and 8 from
the intact carbon chain. Surprisingly, 2-mercapto-3-pentanone
was not detected. Reactions of ribose and cysteine in an
autoclave during 30 min at 140 °C (30) and 20 min up to 145
°C (8) both revealed 3-mercapto-2-pentanone and 2-mercapto-
3-pentanone, with 3-mercapto-2-pentanone dominating. Higher
temperatures and shorter reaction times seem to favor a different
reaction mechanism leading to both isomers. In the present study
the formation of 3-mercapto-2-pentanone via 2,3-pentanedione
and hydrogen sulfide as intermediates (15, 16) apparently does
not play an important role; otherwise, 2-mercapto-3-pentanone
would have been formed at the same time.
To explain the formation of 3-mercapto-2-pentanone without
the simultaneous generation of 2-mercapto-3-pentanone, 5-hy-
droxy-3-mercapto-3-pentene-2-one (19) is proposed as an
intermediate. This intermediate could be formed from the open
chain or cyclic form of the 1,4-dideoxyosone (16) of ribose by
the addition of hydrogen sulfide, originating from cysteine, and
loss of water, as illustrated in Figure 2. 5-Hydroxy-3-mercapto-
3-penten-2-one (19) would be transformed to 3-mercapto-2-
pentanone (8) via a series of reaction steps including a Strecker-
type reaction with cysteine, isomerization, hydrolysis, and
reduction. Its 2-mercapto-3-pentanone isomer is not generated
Figure 2. Proposed formation pathway for 3-mercapto-2-pentanone and
2-methyl-3-furanthiol from ribose and cysteine via the 1,4-dideoxyosone
route.
by this mechanism. On the other hand, dehydration of 2,5-
dihydro-2-hydroxy-2-methyl-3-furanthiol (20), the cyclic form
of 5-hydroxy-3-mercapto-3-penten-2-one, should yield 2-methyl-
3-furanthiol (7). Alternatively, 7 could be formed by dehydration
of 2,5-dihydro-2,3-dihydroxy-2-methylfuran (17, cyclic form of
the 1,4-dideoxyosone of ribose), which yields 3-hydroxy-2-
methylfuran (18). The addition of hydrogen sulfide to its keto
form and a loss of water would lead to 2-methyl-3-furanthiol.
Thus, if proven, the formation of 2-methyl-3-furanthiol via the
1,4-dideoxyosone would represent a new alternative pathway
for the formation via 1-deoxyosone and 4-hydroxy-5-methyl-
3(2H)-furanone (31).
The 1,4-dideoxyosone of ribose, 5-hydroxy-2,3-pentanedione,
has been previously identified as an intermediate in the Maillard
reaction between xylose and amino acids from hydrolyzed wheat
protein (32). Various syntheses have been described by de
Kimpe et al. (33, 34). It also occurs in the alga Laurencia
spectabilis (35) and is, therefore, named laurencione. The
formation of 1,4-dideoxyosones has been explained by the
dehydration of 1-deoxyketose obtained either by the reduction

Aroma Compounds from Ribose and Cysteine
J. Agric. Food Chem., Vol. 51, No. 9, 2003 2717
Figure 4. 13C-Labeling positions in 2-methyl-3-(methylthio)furan (* ) 13C).
Hofmann et al. (40) these disulfides can be formed as artifacts
from the corresponding monothiols during extraction with
organic solvents and storage of the extracts. However, solvent
extraction was avoided in the present study, and the samples
were analyzed directly by HS-SPME-GC-MS. According to van
Seeventer (41) disulfide formation is not the cause of 2-methyl-
3-furanthiol instability in aqueous solutions of reacted cysteine/
ribose kept at 50 °C. To verify whether the disulfides represent
true reaction products or are formed as artifacts during SPME,
additional experiments would be necessary.
Figure 3. Proposed formation pathway for 3-thiophenethiol, adapted from
ref 21.
of 1-deoxyosone, for example, via Strecker degradation, or by
its disproportionation (32, 36). It can be also formed from
oligosaccharides by a so-called “peeling off” mechanism (37).
Further compounds formed from ribose without chain cleav-
age are 2-methylfuran (2) and 2-methylthiophene (4). An
unambiguous determination of the isotopomer ratio for 2-methyl-
3-thiophenethiol (12), the thiophene analogue of 2-methyl-3-
furanthiol, was not possible because it coeluted with another
compound.
For thiazole (3), a molecular ion that was 56% unlabeled (m/z
85) and 44% singly labeled (m/z 86) was found. The signal at
m/z 58 reveals the loss of m/z 27 and 28, corresponding to a
loss of hydrogen [12C]cyanide and hydrogen [13C]cyanide,
respectively. The labeling position is assigned to the C-2 atom
of thiazole, suggesting that it has been formed from cysteamine
derived from cysteine and formaldehyde stemming from ribose
to yield thiazolidine and thiazole by subsequent oxidation of
the former, as proposed by Sakaguchi and Shibamoto (38).
Almost no labeling was observed in 3-thiophenethiol (11),
suggesting its formation from cysteine rather than from ribose.
This is in accordance with the finding that 3-thiophenethiol is
a thermal degradation product of cysteine (39) and that in model
reactions with 4-hydroxy-5-methyl-3(2H)-furanone it occurred
only in cysteine-containing systems (21). Shu et al. (39)
postulated the formation of this compound from two molecules
of mercaptoacetaldehyde, the Strecker aldehyde of cysteine.
Figure 3 gives an overview of the pathway.
An interesting labeling pattern was found for 2-methyl-3-
(methylthio)furan (10). The ratio between the unlabeled, singly
labeled, and the 5- and 6-times-13C-labeled molecules was found
to be 3:1:3:1, suggesting, according to the mass spectra (not
shown), the labeling positions shown in Figure 4. Obviously,
the 2-methylfuran moiety stems exclusively from ribose with
the C-skeleton remaining intact as indicated by the 1:1 ratio
between the unlabeled and the 5-times-labeled molecules. On
the other hand, only part of the carbon atoms of the methylthio
group originate from ribose; otherwise, the labeling pattern
would have been 1:1:1:1. It can be concluded that half of the
thiomethyl carbon atoms stem from cysteine, the reaction
mechanism being still unclear.
In addition to the mercapto compounds 2-methyl-3-furanthiol
and 2-furfurylthiol, the corresponding disulfides bis(2-methyl-
3-furyl) disulfide (13), bis(2-furfuryl) disulfide (15), and the
mixed disulfide between 2-methyl-3-furanthiol and 3-mercapto-
2-pentanone [(2-methyl-3-furyl-2-oxo-3-pentyl) disulfide (14)]
were detected. The isotopomer ratio of approximately 1:2:1
(unlabeled/[13C5]/[13C10]) indicates that they are formed by
oxidation of the corresponding thiols 7-9 as these thiols are
either unlabeled or completely 13C5-labeled. According to
Reaction between 4-Hydroxy-5-methyl-3(2H)-furanone
and Cysteine. 4-Hydroxy-5-methyl-3(2H)-furanone and cysteine
were reacted in a molar ratio of 3:2 using the same reaction
conditions as for the ribose/cysteine experiment. Among other
sulfur compounds 3-mercapto-2-butanone, 2-methyl-3-furanthi-
ol, 3-mercapto-2-pentanone, 2-mercapto-3-pentanone, and bis-
(2-methyl-3-furyl) disulfide were identified. Because both
2-mercapto-3-pentanone and 3-mercapto-2-pentanone are formed
during the reaction, the reaction pathway via the R-diketones,
as depicted in Figure 5, is postulated. The addition of the amino
group of cysteine causes ring opening of 4-hydroxy-5-methyl-
3(2H)-furanone (21). The resulting Schiff base of the 1-deoxy-
pentosone with cysteine undergoes Strecker degradation and
yields 2-amino-4,5-dihydroxy-3-pentanone (22). The retroaldol
cleavage of the compound forms 2-amino-4-hydroxy-3-butanone
(23), which by a series of reaction steps, including enolization,
dehydration, and hydrolysis, is converted to 2,3-butandione (24).
This compound gives 3-mercapto-2-butanone (5) upon reaction
with hydrogen sulfide and reduction. If, instead of a retroaldol
cleavage, 22 loses ammonia, 1-hydroxy-2,3-pentadione (25) is
formed. Strecker degradation with cysteine and loss of water
yield 2,3-pentadione (26). Whitfield and Mottram (21) postulate
a similar mechanism that, however, does not involve cysteine.
If cysteine is present, the authors propose a mechanism including
an aldol reaction between 2-oxopropanal, which is generated
by the fragmentation of 4-hydroxy-5-methyl-3(2H)-furanone,
and acetaldehyde stemming from cysteine degradation. CAMO-
LA experiments with 4-hydroxy-5-methyl-3(2H)-furanone/cys-
teine would elucidate whether R-mercaptoketones are formed
from an intact 4-hydroxy-5-methyl-3(2H)-furanone chain or
from fragments from 4-hydroxy-5-methyl-3(2H)-furanone and
cysteine. Further experiments are currently under way to answer
this question.
Reaction between 4-Hydroxy-5-methyl-3(2H)-furanone,
[13C5]Ribose, and Cysteine. In this experiment, [13C5]ribose
and cysteine were reacted with 4-hydroxy-5-methyl-3(2H)-
furanone. Because [13C5]ribose and 4-hydroxy-5-methyl-3(2H)-
furanone carried different carbon isomer labels, it was possible
to estimate the contribution of ribose and 4-hydroxy-5-methyl-
3(2H)-furanone to the formation of aroma compounds. Table
3 shows the identified compounds, the number of 13C-labeled
atoms in the molecules, and the ratios of the isotopomers. The
two mercaptoketones 3-mercapto-2-butanone and 2-mercapto-
3-pentanone were mainly unlabeled (94 and 96%, respectively)
and, hence, almost exclusively formed from 4-hydroxy-5-
methyl-3(2H)-furanone, whereas 3-mercapto-2-pentanone (42%
unlabeled and 58% 13C5-labeled) was formed from both 4-hy-

2718 J. Agric. Food Chem., Vol. 51, No. 9, 2003
Cerny and Davidek
Figure 5. Proposed formation pathway for R-mercaptoketones from 4-hydroxy-5-methyl-3(2H)-furanone.
Table 3. Proportion of Labeling in Compounds from the Reaction between 4-Hydroxy-5-methyl-3(2H)-furanone, [13C ]Ribose, and Cysteine
5
no. of 13C atoms in
the labeled molecule
unlabeled
compound (%)
13C-labeled
compound (%)
+
no.
compounda
m/z (M )
5
6
3-mercapto-2-butanone
2-furaldehyde
104
96
5
5
94
2
6
98
7
8
2-methyl-3-furanthiol
3-mercapto-2-pentanone
2-mercapto-3-pentanone
2-furfurylthiol
2-methyl-3-(methylthio)furan
3-thiophenethiol
114
118
118
114
128
116
226
5
5
5
5
5, 6
4
5, 10
7
93
58
4
100
84, 10
2
42
96
0
6
98
1
16
9
10
11
13
bis(2-methyl-3-furyl) disulfide
15, 84
a Compounds were identified in the corresponding reaction between ribose and cysteine by comparing the mass spectra and retention indices with those of authentic
compounds analyzed in our laboratory or in literature data.
droxy-5-methyl-3(2H)-furanone and ribose. Both reaction path-
ways shown in Figures 2 and 5 seem to contribute to the
formation of 3-mercapto-2-pentanone. These findings, however,
do not allow conclusions on the role of 4-hydroxy-5-methyl-
3(2H)-furanone as intermediate in the formation of the mer-
captoketones 3-mercapto-2-butanone and 2-mercapto-3-pen-
tanone from ribose and cysteine alone. 2-Furaldehyde and
2-furfurylthiol, which were ∼100% labeled, stem from ribose.
of 1-deoxypentosone from 4-hydroxy-5-methyl-3(2H)-furanone
is not favored under the conditions studied or 1,4-dideoxypen-
tosone arises from ribose by a mechanism other than the
reduction or disproportionation of 1-deoxypentosone. Otherwise,
2-methyl-3-furanthiol should be formed also from 4-hydroxy-
5-methyl-3(2H)-furanone.
The alternative pathways leading to 1,4-dideoxypentosone
from ribose are proposed in Figure 6. The reaction of ribose
with cysteine leads via Amadori rearrangement to 1-cysteino-
1,4-dideoxypentosone (28). The latter compound is transformed
via Strecker degradation and hydrolysis to 1,4-dideoxypentosone
(16). This mechanism is analogous to that proposed by Nedvidek
et al. (32) for the formation of 1,4-dideoxyosones from
1-deoxyosones. Recently, the high mobility of the carbonyl
group in 1-amino-1,4-dideoxyosones has been demonstrated
(42). Thus, 1,4-dideoxypentosone could also be formed from
1-cysteino-1,4-dideoxypentosone by enolizations yielding the
Schiff base of 4-deoxypentosone with cysteine (29). Strecker
degradation of the latter compound followed by elimination of
ammonia would yield 1,4-dideoxypentosone. Further experi-
ments are necessary to verify whether the new pathway proposed
Surprisingly, 2-methyl-3-furanthiol was mostly labeled (93%),
indicating that it originates mainly from ribose and only to a
small extent from 4-hydroxy-5-methyl-3(2H)-furanone. Also,
most of the corresponding disulfide 13 was found fully labeled.
These findings suggest that 4-hydroxy-5-methyl-3(2H)-furanone
plays only a minor role as an intermediate in the formation of
2-methyl-3-furanthiol from ribose and cysteine under the
conditions studied. 2-Methyl-3-furanthiol formation from ribose
by recombination of fragments can be excluded, too, according
to the results from the first experiment.
As postulated in Figure 2, 1,4-dideoxypentosone could be a
suitable precursor of 2-methyl-3-furanthiol. However, if this
mechanism is correct, the results imply that either the formation

Aroma Compounds from Ribose and Cysteine
J. Agric. Food Chem., Vol. 51, No. 9, 2003 2719
Figure 6. Proposed formation of 1,4-dideoxypentosone from ribose and cysteine.
in Figure 2 can be considered to be the main reaction
mechanism and the 1,4-dideoxyosone as a new key intermediate
for 2-methyl-3-furanthiol formation in the reaction of ribose and
cysteine under cooking conditions.
3-furanthiol under cooking conditions. A new reaction mech-
anism with the 1,4-dideoxyosone as key intermediate is
proposed. In the reaction of ribose, 4-hydroxy-5-methyl-3(2H)-
furanone, and cysteine, 2-mercapto-3-pentanone was formed
almost exclusively from 4-hydroxy-5-methyl-3(2H)-furanone,
whereas 3-mercapto-2-pentanone was formed from both 4-hy-
droxy-5-methyl-3(2H)-furanone and ribose. These data suggest
that the formation of 3-mercapto-2-pentanone from ribose by
an alternative pathway is more important than its formation via
4-hydroxy-5-methyl-3(2H)-furanone under the conditions stud-
ied. The pathway involving the addition of hydrogen sulfide to
the cyclic form of the 1,4-dideoxyosone explains why 3-mer-
capto-2-pentanone rather than 2-mercapto-3-pentanone is formed
from ribose in the Maillard reaction with cysteine.
Reaction of 2-Furaldehyde, [13C5]Ribose, and Cysteine.
2-Furaldehyde was reacted with [13C5]ribose and cysteine to
elucidate its role as a 2-furfurylthiol precursor. During HS-
SPME-GC-MS analysis only a few compounds were detected
due to saturation of the PDMS-DVB fiber with unreacted
2-furaldehyde. However, 2-furfurylthiol was identified, and the
ratio between unlabeled and 13C5-labeled isotopomer was found
to be 92:8 according to the ratio between the ion signals m/z
114 and 119. This indicates that most 2-furfurylthiol was formed
from 2-furaldehyde. Because ribose and 2-furaldehyde were
employed in equimolar amounts in the reaction, it can be
concluded that 2-furaldehyde is a much more efficient precursor
than ribose. In contrast to 4-hydroxy-5-methyl-3(2H)-furanone,
its role as an intermediate for 2-methyl-3-furanthiol formation
under the conditions studied being doubtful, this experiment
confirmed the role of 2-furaldehyde as an important intermediate
for 2-furfurylthiol generation during the Maillard reaction
between ribose and cysteine.
ABBREVIATIONS USED
CAMOLA, carbohydrate module labeling; HS-SPME-GC-
MS, headspace solid-phase microextraction-gas chromatogra-
phy-mass spectrometry; PDMS-DVB, poly(dimethylsiloxane)-
divinylbenzene; TIC, total ion count.
ACKNOWLEDGMENT
Conclusions. The reaction between ribose and cysteine in
aqueous phosphate buffer (pH 5) generates numerous sulfur-
containing aroma compounds. The ribose carbon chain remains
intact in the reaction yielding 2-methyl-3-furanthiol, 2-furfu-
rylthiol, and 3-mercapto-2-pentanone. 2-Furaldehyde could be
confirmed as an important intermediate for 2-furfurylthiol
formation. On the other hand, 4-hydroxy-5-methyl-3(2H)-
furanone plays only a minor role in the formation of 2-methyl-
We thank Tuong Huynh-Ba for valuable discussions and
Elizabeth Prior and Hedwig Schlichtherle-Cerny for critically
reading the manuscript.
LITERATURE CITED
(1) Morton, I. D.; Akroyd, P.; May, C. G. Flavoring substances.
U.S. Patent 2934437, 1960.

2720 J. Agric. Food Chem., Vol. 51, No. 9, 2003
Cerny and Davidek
(2) Gasser, U.; Grosch, W. Identification of volatile flavor com-
pounds with high aroma values from cooked beef. Z. Lebensm.
Unters. Forsch. 1988, 186, 489-494.
(3) Evers, W. J.; Heinsohn, H. H., Jr.; Mayers, B. J.; Sanderson, A.
Furans substituted at the three position with sulfur. In Phenolic,
Sulfur and Nitrogen Compounds in Food FlaVors; Charalambous,
G., Katz, I., Eds.; American Chemical Society: Washington, DC,
1976; pp 184-193.
(4) Gasser, U.; Grosch, W. Primary odorants of chicken broth. A
comparative study with meat broths from cow and ox. Z.
Lebensm. Unters. Forsch. 1990, 190, 3-8.
(5) Gasser, U.; Grosch, W. Aroma extract dilution analysis of
commercial meat flavorings. Z. Lebensm. Unters. Forsch. 1990,
190, 511-515.
(6) Farmer, L. J.; Mottram, D. S. Recent studies on the formation
of meat-like aroma compounds. In FlaVour Science and Tech-
nology; Bessiere, Y., Thomas, A. F., Eds.; Wiley: Chichester,
U.K., 1990; pp 113-116.
(7) Farmer, L. J.; Mottram, D. S.; Whitfield, F. B. Volatile
compounds produced in Maillard reactions involving cysteine,
ribose and phospholipid. J. Sci. Food Agric. 1989, 49, 347-
368.
(21) Whitfield, F. B.; Mottram, D. S. Investigation of the reaction
between 4-hydroxy-5-methyl-3(2H)-furanone and cysteine or
hydrogen sulfide at pH 4.5. J. Agric. Food Chem. 1999, 47,
1626-1634.
(22) Schieberle, P. Quantitative analysis of aroma compounds in wheat
and rye bread crusts using a stable isotope dilution assay. J.
Agric. Food Chem. 1987, 35, 252-257.
(23) Koehler, P. E.; Mason, M. E.; Newell, J. A. Formation of
pyrazine compounds in sugar-amino acid model systems. J.
Agric. Food Chem. 1969, 17, 393-396.
(24) Hofmann, T.; Schieberle, P. Acetylformoinsan important pro-
genitor of 4-hydroxy-2,5-dimethyl-3(2H)-furanone and 2-acetyltet-
rahydropyridine during thermal food processing. In FlaVour
2000; Rothe, M., Ed.; Eigenverlag: Bergholz-Rehbru¨cke, Ger-
many, 2001; pp 311-322.
(25) Schieberle, P.; Bareth, A.; Fischer, R.; Hofmann, T. The
carbohydrate modul labeling-techniquesA useful tool to clarify
formation pathways of key aroma and taste compounds formed
in Maillard-type reactions. Book of Abstracts, 10th Weurman
Flavour Research Symposium, Beaune, France, June 24-28;
INRA: Dijon, France, 2002; Abstract O25.
(26) Frank, O.; Hofmann, T. Reinvestigation of the chemical structure
of bitter-tasting quinizolate and homoquinizolate and their
Maillard-type formation pathways using suitable 13C-labeling
experiments. J. Agric. Food Chem. 2002, 50, 6027-6036.
(27) Gry, J. Guidelines for Safety EValuation of Thermal Process
FlaVourings; Council of Europe Publishing: Strasbourg, France,
1995.
(28) Hofmann, T.; Schieberle, P. Quantitative model studies on the
effectiveness of different precursor systems in the formation of
the intense food odorants 2-furfurylthiol and 2-methyl-3-
furanthiol. J. Agric. Food Chem. 1998, 46, 235-241.
(29) Silwar, R.; Tressl, R. Gas chromatographic-mass spectrometric
investigation of aroma compounds formed in the cysteine-
methionine-furfural model system under roasting conditions.
Z. Lebensm. Unters. Forsch. 1989, 189, 205-211.
(8) Hofmann, T.; Schieberle, P. Evaluation of the key odorants in a
thermally treated solution of ribose and cysteine by aroma extract
dilution techniques. J. Agric. Food Chem. 1995, 43, 2187-2194.
(9) Mulders, E. J. Volatile components from the nonenzymic
browning reaction of the cysteine/cystine-ribose system. Z.
Lebensm. Unters. Forsch. 1973, 152, 193-201.
(10) Whitfield, F. B.; Mottram, D. S.; Brock, S.; Puckey, D. J.; Salter,
L. J. Effect of phospholipid on the formation of volatile
heterocyclic compounds in heated aqueous solutions of amino
acids and ribose. J. Sci. Food Agric. 1988, 42, 261-272.
(11) Guentert, M.; Bertram, H. J.; Hopp, R.; Silberzahn, W.; Sommer,
H.; Werkhoff, P. Thermal generation of flavor compounds from
thiamin and various amino acids. In Recent DeVelopments in
FlaVor and Fragrance Chemistry; Hopp, R., Mori, K., Eds.;
VCH: Weinheim, Germany, 1993; pp 215-240.
(12) Misharina, T. A.; Golovnya, R. V. Sulfur-containing compounds
in meat aroma model systems. Abh. Akad. Wiss. DDR, Abt.
Math., Naturwiss., Tech. 1988, 115-127.
(13) Grosch, W.; Zeiler-Hilgart, G.; Cerny, C.; Guth, H. Studies on
the formation of odorants contributing to meat flavors. In
Progress in FlaVor Precursors Studies; Schreier, P., Winterhalter,
P., Eds.; Allured Publishing: Carol Stream, IL, 1993; pp 329-
342.
(14) Mottram, D. S.; Nobrega, I. C. C. Formation of sulfur aroma
compounds in reaction mixtures containing cysteine and three
different forms of ribose. J. Agric. Food Chem. 2002, 50, 4080-
4086.
(15) Hofmann, T. Characterization of intense odorants in carbohydrate/
cysteine model reactions and elucidation of formation pathways.
Ph.D Dissertation, Technical University Munich, 1995 (in
German).
(16) Mottram, D. S.; Madruga, M. S.; Whitfield, F. B. Some novel
meatlike aroma compounds from the reactions of alkanediones
with hydrogen sulfide and furanthiols. J. Agric. Food Chem.
1995, 43, 189-193.
(17) Shibamoto, T. Formation of sulfur- and nitrogen-containing
compounds from the reaction of furfural with hydrogen sulfide
and ammonia. J. Agric. Food Chem. 1977, 25, 206-208.
(18) Severin, T.; Seilmeier, W. Maillard reaction. II. Reaction of
pentoses in the presence of amine acetates. Z. Lebensm. Unters.
Forsch. 1967, 134, 230-232 (in German).
(19) Van den Ouweland, G. A. M. Components contributing to beef
flavor. Volatile compounds produced by the reaction of 4-hy-
droxy-5-methyl-3(2H)-furanone and its thio analog with hydro-
gen sulfide. J. Agric. Food Chem. 1975, 23, 501-505.
(20) Whitfield, F.; Mottram, D. S. Heterocyclic volatiles formed by
heating cysteine or hydrogen sulfide with 4-hydroxy-5-methyl-
3(2H)-furanone at pH 6.5. J. Agric. Food Chem. 2001, 49, 816-
822.
(30) Nobrega, I. C. C.; Mottram, D. S. The formation of meaty aroma
compounds in cysteine model systems containing ribose 5-phos-
phate or ribose. In Frontiers of FlaVour Science; Schieberle, P.,
Engel, K. H., Eds.; DFA: Garching, Germany, 2000; pp 487-
491.
(31) Mottram, D. S. Flavour formation in meat and meat products:
a review. Food Chem. 1998, 62, 415-424.
(32) Nedvidek, W.; Ledl, F.; Fischer, P. Detection of 5-hydroxy-
methyl-2-methyl-3(2H)-furanone and of R-dicarbonyl compounds
in reaction mixtures of hexoses and pentoses with different
amines. Z. Lebensm. Unters. Forsch. 1992, 194, 222-228.
(33) De Kimpe, N.; Georgieva, A.; Boeykens, M.; Lazar, L. Synthesis
of laurencione, a labile dihydro-3(2H)-furanone derivative from
the red alga Laurencia spectabilis. J. Org. Chem. 1995, 60,
5262-5265.
(34) De Kimpe, N.; Georgieva, A.; Boeykens, M.; Kozekov, I.;
Aelterman, W. New formal syntheses of laurencione, a labile
dihydrofuranone derivative from the red alga Laurencia spec-
tabilis. Synthesis 1995, 1131-1134.
(35) Bernart, M. W.; Gerwick, W. H.; Corcoran, E. E.; Lee, A. Y.;
Clardy, J. Laurencione, a heterocycle from the red alga Laurencia
spectabilis. Phytochemistry 1992, 31, 1273-1276.
(36) Glomb, M. A.; Pfahler, C.; Hiller, R. Reductones participate in
redox-reactions during amine catalyzed sugar degradation. In The
Maillard Reaction in Food Chemistry and Medical Science:
Update for The Postgenomic Era; Horiuchi, S., Taniguchi, N.,
Hayase, F., Kurata, T., Osawa, T., Eds.; Elsevier Science:
Amsterdam, The Netherlands, 2002; pp 201-202 (Japan Maillard
Reaction Society: Kumamoto, Japan, 2001; Abstract F01-002).
(37) Hollnagel, A.; Krol, L. W. Degradation of oligosaccharides in
nonenzymatic browning by formation of R-dicarbonyl com-
pounds via a “peeling off” mechanism. J. Agric. Food Chem.
2000, 48, 6219-6226.

Aroma Compounds from Ribose and Cysteine
J. Agric. Food Chem., Vol. 51, No. 9, 2003 2721
(38) Sakaguchi, M.; Shibamoto, T. Formation of heterocyclic com-
pounds from the reaction of cysteamine and D-glucose, acetal-
dehyde, or glyoxal. J. Agric. Food Chem. 1978, 26, 1179-1183.
(39) Shu, C.-K.; Hagedorn, M. L.; Mookherjee, B. D.; Ho, C.-T. pH
effect on the volatile components in the thermal degradation of
cysteine. J. Agric. Food Chem. 1985, 33, 442-446.
of thiols in an aqueous process flavoring. J. Agric. Food Chem.
2001, 49, 4292-4295.
(42) Biemel, K. M.; Conrad, J.; Lederer, M. O. Unexpected carbonyl
mobility in aminoketose: the key to major Maillard crosslinks.
Angew. Chem., Int. Ed. Engl. 2002, 41, 801-804.
(40) Hofmann, T.; Schieberle, P.; Grosch, W. Model studies on the
oxidative stability of odor-active thiols occurring in food flavors.
J. Agric. Food Chem. 1996, 44, 251-255.
Received for review November 13, 2002. Revised manuscript received
February 21, 2003. Accepted March 2, 2003.
(41) Van Seeventer, P. B.; Weenen, H.; Winkel, C.; Kerler, J. Stability
JF026123F